Endometrioid endometrial adenocarcinoma is the most common gynecologic cancer in the United States, which can be prevented by early diagnosis and surgical or hormonal ablation of premalignant lesions. Poor reproducibility of precancer diagnosis by pathologists, and paucity of controlled prospective therapeutic studies have hindered uniform and effective early intervention. This project is a laboratory translational component to Gynecologic Oncology Group ongoing clinical trial (G0G167) which enrolls patients with suspected endometrial precancers into an immediate hysterectomy arm (untreated, Part A), or progestin therapy arm followed by hysterectomy (hormone treated, Part B). The first two Aims will establish the clinical cancer-predictive value of endometrial precancers diagnosed by PTEN immunohistochemistry (SA#1), or computerized morphometry of routinely stained slides (SA#2). PTEN is a tumor suppressor gene altered very early in endometrial carcinogenesis which displays decreased protein expression in 75 percent of premalignant and malignant endometrioid endometrial lesions, thereby demarcating them with unprecedented physical resolution (single glands). We will develop a quantitative scale for classifying size and complexity of premalignant PTEN-null endometrial lesions, thereby stratifying them as low or high risk. Computerized image analysis of routinely stained (hematoxylin and eosin) endometrial tissues and D-Score calculation will also be used to identify a class of premalignant lesions in the untreated arm (Part A) of G0G167. One of the two methods will be identified by its cancer predictive value as a preferred means of diagnosing high risk endometrial precancers. The third Aim will use one of the methods from SAl & 2 to identify a group of patients with high-risk precancers, and measure regression of those premalignant lesions in the resected uterus. Comparison of regression rates of untreated patients with those of progesterone treated patients will allow us to define efficacy of these hormonal regimens as alternatives to surgical hysterectomy. A major advantage of this proposal is that its well defined clinical setting provides contextual relevance to promising highly sensitive and reproducible methods of endometrial precancer diagnosis. If successful, we will be able to rapidly implement a rational and effective strategy for early detection and chemoprevention of endometrial carcinoma.